[TriLUG] Why the message size limit? [Was: Your message to TriLUG awaits moderator approval]

[Cristóbal Palmer via TriLUG]

On 6/7/18 10:34 AM, Matt Flyer via TriLUG wrote:
> Let's ask a couple of practical questions.
> 
> What is the purpose of the size limit?  Is it routinely stopping abuse
> or otherwise serving a real requirement?  Or is it an impediment to the
> legitimate mission of this list and organization?

I was around at the time this was first implemented. In my memory, I was
not a terribly active participant in the conversation about what to do,
but I do remember the broad strokes. What I say now is from that
(fallible) memory, so if anybody else who is/was on the SC or feels like
digging through the steering@ archives wants to correct me when/where
I'm wrong, that would be great.

There are two rationales for the limit, and they both involve mailman.

1. Archives

This rationale is about being able to go back to the archives and (a)
follow a thread easily, and (b) not have to over-invest in storage for
the archives. Additionally, the archiver (pipermail) that we were using
at the time of the decision is Not Smart, and would do...
"interesting"... things with big MIME emails (eg. just report that a
MIME part was missing and have a node in the thread tree be missing for
a viewer of the archives).

2. Digest readers

This rationale is about enforced courtesy to those list subscribers who
read the digest version produced by mailman. Since repeated reminders
asking people to trim replies and/or avoid top-posting were ignored (and
even mocked), this argument holds that if we provide a digest at all, we
have a duty to make sure it's at least somewhat readable to digest
readers. Threads with many replies would rapidly grow to be mostly
signatures and "on [date], [person] wrote" lines that were not necessary
to parse the reply. In my memory, Kevin Hunter did the work to identify
what percentage of messages that were _not_ this sort of reply chain
would be caught at different thresholds.

That's my memory. This is not me making an argument to keep the limit or
to keep it at its current value. This is just an effort by a person who
has been on this list for... a long while... to provide some
institutional memory that can help folks in making informed decisions.
To the extent that I want to make any argument for any position, my
argument is this: ask for people who currently have sufficient access to
revisit this by collecting data again (eg. how many people are digest
subscribers) instead of jumping directly to any one technical change and
arguing for it. This list serves multiple audiences consuming it in
multiple formats on multiple time scales. I would encourage folks to
gather facts and reflect on how we can encourage the best possible
stewardship for as many folks as possible. I'll be even more bold:
please volunteer to help research a new MLM (newer mailman? other?),
archiver, and other list management tools. Do requirements gathering and
other project management before assuming that any change(s) will make
things better.

Cheers,
--
Cristóbal Palmer